lovedheart/Qwen3-Coder-Next-REAP-40B-A3B
LoRA-style (fewer layers) vs Qwen/Qwen3-Coder-Next. lovedheart/Qwen3-Coder-Next-REAP-40B-A3B matches part of Qwen/Qwen3-Coder-Next (37479 of its 74391 tensors; the rest of Qwen/Qwen3-Coder-Next has no counterpart). On the shared part: Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, up_proj changed, in 48 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/157, Attention 0/264, MLP 1176/1372, lm_head 0/1
4 derivatives checked
- mradermacher/Qwen3-Coder-Next-REAP-40B-A3B-i1-GGUF: Quantization
- mradermacher/Qwen3-Coder-Next-REAP-40B-A3B-GGUF: Quantization
- RepublicOfKorokke/Qwen3-Coder-Next-REAP-40B-A3B-mlx-mxfp4: Quantization
- RepublicOfKorokke/Qwen3-Coder-Next-REAP-40B-A3B-mlx-nvfp4: Quantization